SUBJECT: &NAME : &NAME &NAME Tuesday 11th March Natural Language and Information Processing Group Seminar &WEBSITE 12pm , Tuesday 11th March , FW11 &ORG , &NAME &NAME Building , &NAME &NAME Avenue &WEBSITE &NAME : Experiments on the Automatic Induction of German Verb Classes Speaker : &NAME &NAME im &NAME - University of &NAME &NAME : The verb is central to the structure and the meaning of a sentence , so available lexical resources on verb knowledge represent an important means for Natural Language Processing applications . This talk focuses on the automatic induction of verb classes : a clustering methodology is developed which utilises the relationship between verb meaning components and verb behaviour to provide a classification of German verbs . In the first part of the talk , I describe the acquisition of lexical syntactic and semantic verb properties from an unsupervised trained statistical grammar model . The properties are used to define German verbs on &NUM levels of verb description , purely syntactic frame types , prepositional phrase information and selectional preferences . In the second part of the talk , I describe the adjustment of the standard clustering algorithm k-Means to the task of inducing semantic verb classes from the various verb descriptions . The clustering parameters , clustering evaluation and clustering results are interpreted and discussed . ( All welcome - For instructions on how to get to the &ORG , please see : &WEBSITE